The general objective of this project is to evaluate consumption measures derived from dietary data in epidemiologic studies of alcohol as a risk factor for chronic disease. Four existing data sets will be analyzed: the first and second National Health and Nutrition Examination Surveys (NHANES I and II); the Nationwide Food Consumption Survey (NFCS); and data from the Food Frequency study carried out under contract to the U.S. Dep't of Agriculture. In each of these surveys, information about alcohol consumption was collected for the same individuals by at least two different methods, including 24-hour dietary recalls, food frequency questionnaires, food diaries and a modified quantity-frequency questionnaire. This project is intended to evaluate the overall strength and weaknesses of dietary measures of alcohol consumption, to compare the results from different methods and to explore ways in which to make the most efficient use of the alcohol information from dietary data. The reliability and relative validity of alcohol consumption estimates from such data will be assessed by comparing the estimates from different methods to each other within a survey, by comparing these estimates to physiological measurements known to be related to alcohol consumption and by comparing aggregate estimates from national dietary surveys with estimates from national surveys of drinking behavior. The dietary data will also be analyzed to describe certain aspects of the epidemiology of alcohol consumption, including differences in consumption pattern of different types of alcoholic beverages, dietary intake associated with alcohol consumption and patterns of alcohol consumption including seasonal variation and intra-individual variation. Methods of assessing the potential bias and loss of power associated with imperfect measurement of risk factors will be used, in conjunction with the findings from these data analyses, to evaluate measures of alcohol consumption derived from dietary data and their appropriate uses in epidemiologic research.